Anatomical structures in the maxillary sinus related to lateral sinus elevation: a cone beam computed tomographic analysis

Authors


Corresponding author:

Dr Jong-Hyuk Chung

Department of Periodontology, Institute of Oral Biology, School of Dentistry, Kyung Hee University, 1 Hoegi-dong, Dongdaemun-gu, Seoul 130-701, Korea

Tel.: 82 2 958 9382

Fax: 82 2 958 9387

e-mail: chungjh@khu.ac.kr

Abstract

Objectives

The objective of this study is to evaluate the anatomical structures in the maxillary sinus with relation to lateral approach sinus elevation utilizing cone beam computed tomography (CT) scans taken prior to sinus elevation surgery.

Materials and methods

A total of 150 CT images were acquired from 150 patients (90 men and 60 women; mean age, 49.4 years, range 23–86 years) who were being treated with implant-supported restorations in the posterior edentulous maxilla. Of the 150 CT scans, 65 were of the right sinus and 85 of the left sinus. Measurements of the anatomical structures in the maxillary sinus were conducted on the CT images.

Results

In the mean width of the lateral wall, there were statistically significant values among the measurement points (< 0.05). The anterior area of the sinus lateral wall was thicker than the posterior lateral wall. There was a statistically significant difference between the vessel diameter and lateral wall width (< 0.05). As sinus lateral wall width increased, so did the vessel diameter. The mean distance to the inferior border of the vessel from the sinus floor and from the alveolar crest was 8.25 and 17.03 mm, respectively. The intraosseous group among the vessel position was 64.3%, so the intraosseous vessel could be visualized in CT scans at 64.3%. In angle A, the group of less than 30° was 4.8%. Schneiderian membrane perforation by narrow angle had a low risk. The prevalence of the septa related to Schneiderian membrane perforation was 44%. The distance to the inferior border of the vessel from the alveolar crest being less than 15 mm was 31%. The vessel diameter greater than 1 mm was 37.8%.

Conclusions

Based on present research about utilizing cone beam CT scans for sinus elevation, the alteration of the lateral approach sinus elevation technique is highly recommended if complications such as membrane perforation or bleeding are expected.

The maxillary posterior region is often a problem area for the placement of implants. In edentulous patients, the alveolar bone is atrophied and may present a paper-thin bone wall on the lateral and occlusal sides (Garg 1999; van den Bergh et al. 2000). Reduction of the posterior maxilla seems to be influenced by the duration of edentulousness, as well as by osteoporotic changes, without being directly related to ridge configuration and bone volume (Ulm et al. 1995a). Thus, the purpose of sinus elevation surgery is to restore a sufficient amount of alveolar bone so that implants can be successfully placed. The method for elevation of the Schneiderian membrane for the augmentation of the maxillary sinus was first published by Boyne & James (1980). A variety of grafting materials and surgical techniques have been applied for sinus elevation (Fugazzotto & Vlassis 1998; Jensen et al. 1998).

The maxillary sinus is a quadrangular pyramid that has an internal (lingual) base and an apex extending into the zygomatic process of the maxilla (Chanavaz 1990). The maxillary sinus is a 15-ml-volume airspace (Garg 1999). The septa may divide the sinus into two or more cavities that may or may not communicate with each other (Underwood 1910). The bony walls of the sinus are thin, except for the anterior wall and alveolar ridge in dentate individuals (Garg 1999).

The Schneiderian membrane, attached to the bordering bone of the maxillary sinus and characterized by a periosteum overlaid with a thin layer of pseudociliated stratified respiratory epithelium, constitutes an important barrier for the protection and defense of the sinus cavity (Smiler et al. 1992).

The blood supply of the maxillary sinus derives from the infraorbital artery (IOA), the greater palatine artery, and the posterior superior alveolar artery (PSAA) (Chanavaz 1990; Solar et al. 1999) (Fig. 1). According to Solar et al., several anastomoses of the PSAA and the IOA can be found usually inside the bony lateral wall, which supplies the Schneiderian membrane as well as the epiperiosteal tissues. Anatomically, an anastomosis between the PSAA and IOA is always found at the lateral wall (Solar et al. 1999; Traxler et al. 1999).

Figure 1.

Schematic diagram of the left maxilla, lateral view, with blood vessel. An anastomosis between the PSAA and IOA is found at the lateral wall. The lateral wall with zygomatic process has been removed in this diagram.

Sinus elevation is not associated with an increased risk for implant failure (Schwartz-Arad et al. 2004; McDermott et al. 2006). However, it can be complicated because of intrasinusal anatomic structures (Garg 1999; Flanagan 2005). Perforation of the Schneiderian membrane is the main intraoperative complication and occurs in 11% to 56% of procedures according to a review of the literature (Testori et al. 2008). Anatomical and technical factors have been implicated in membrane perforation (Ulm et al. 1995b). A second intraoperative complication is the profuse bleeding that occurs while performing an antrostomy with a rotary cutting instrument. This occurs when the anastomosis of the lower branch of the PSAA and IOA is severed, most commonly by vertical osteotomy cutting (Wallace et al. 2007). Therefore, during sinus elevation, several parameters must be considered. Knowledge of the blood supply and other anatomical structures of the maxillary sinus are necessary to avoid unnecessary complications.

Conditions such as sinus floor convolutions, sinus septa, and narrow sinuses can complicate membrane elevation and increase the risk of membrane perforation during the procedure (Chanavaz 1990; Ulm et al. 1995b; Cho et al. 2001; Schwartz-Arad et al. 2004). Perforation is most likely to happen at sharp edges and ridges like the Underwood septa and at the spines (Chanavaz 1990). According to Cho et al., angle A, which is the angle between the buccal alveolar and palatal alveolar walls (Fig. 2h), relates to the mediolateral width of the sinus in that the more acute the angle, the narrower the sinus, and the steeper the medial and lateral walls. Certain surgical procedural modifications can be done to minimize the risk of perforation (Cho et al. 2001).

Figure 2.

CT images. (a) Axial image. “z,” as reference point, is the zygomatic process of the maxilla. “z − 5” is 5 mm of “z” in the front. “z + 5” is 5 mm of “z” in the rear. (b)–(h) Coronal images: (b) thickness of the sinus lateral wall at the areas 3 mm from the sinus floor; (c) thickness of the sinus lateral wall at the areas 13 mm from the sinus floor; (d) thickness of the sinus lateral wall at the areas 15 mm from the alveolar crest; (e) distance to the inferior border of the vessel from the sinus floor; (f) distance to the inferior border of the vessel from the alveolar crest; (g) vessel diameter and position, on this view, the vessel is intraosseous; (h) angle A, on this view, angle A belongs in group 3.

Nevertheless, sinus elevation is efficacious and effective as an adjunctive procedure for implant-supported restorations in posterior maxilla lesions (Jensen et al. 1998). Adequately repaired perforations have no effect on the ultimate survival of the implants placed in the affected sinuses (Testori et al. 2008).

Panoramic radiographs, computed tomography (CT), and sinus radiography help the surgeon determine the presence of antral cloudiness or pathology, antral size, the available maxillary alveolar bone height, the location of sinus floor convolutions (septi), and the surgical entry site (Smiler et al. 1992; Garg 1999). CTs are the most accurate tool to evaluate important anatomic parameters, such as the existence of a large diameter anastomosis, the width of the lateral wall, and pathology of the Schneiderian membrane and sinus septa (Velásquez-Plata et al. 2002; Elian et al. 2005).

In recent years, the blood supply and other anatomical structures of the maxillary sinus using CT scans have been studied by several authors (Velásquez-Plata et al. 2002; Elian et al. 2005; Kim et al. 2006; Mardinger et al. 2007; Ella et al. 2008). Earlier studies reported results that were based on only 50 cases, or measurement points that were from either the alveolar crest or maxillary sinus floor only. Furthermore, the thickness of the sinus lateral wall was not studied.

The purpose of this study is to evaluate the anatomical structures in the maxillary sinus with relation to lateral sinus elevation in Korean adults utilizing cone beam computed tomography (CBCT) scans taken prior to surgery and to discuss its importance.

Material and methods

Materials

Only maxillary CBCT images were included in this study. The sample population consisted of 150 patients (90 men and 60 women, mean age of 49.35 years ranging from 23 and 86 years) for whom treatment was being planned for implant-supported restorations in the posterior edentulous maxilla. Of the 150 CT scans, 65 were of the right sinus and 85 were of the left. The protocol was approved by the Institutional Review Board for Clinical Research in the Kyung Hee Dental Hospital (KHD IRB 2009-4).

Methods

Cone beam computed tomographs

Cone beam computed tomography was performed on a PSR 9000N (Asahi Roentgen Ind. Co., Ltd, Kyoto, Japan). Patients were positioned parallel to the office floor with a Frankfort-Horizontal plane. Sagittal and coronal images were obtained using the dental mode at 80 kV, 10 mA for 13.3 s. The images were obtained from volume data of the cone type with a 40 mm high and 41 mm diameter field size. The sagittal, coronal, and axial images were reformatted using a software program (OnDemand3D, Cybermed Inc., Seoul, Korea). The slice thickness of the multiplanar reconstruction images was 0.1 mm. The measurements were made to the nearest 0.01 mm with a caliper.

Selection of the measurement images

On the axial images, the reference point was determined to be decussate intersection of the most prominent point of the zygomatic process of the maxilla and the line parallel to the centerline of the alveolar ridges. Each area 5 mm, 10 mm, and 15 mm from the reference point was calculated in the front and rear of the images. The reference point were labeled “z” while other points were labeled “z − 5,” “z − 10,” and “z − 15” in the front and “z + 5,” “z + 10,” and “z + 15” in the rear. In this way, a total of seven coronal images per subject were calculated (Fig. 2a).

Measurements

Thickness of the sinus lateral wall

Thickness of the sinus lateral wall 3 mm from the sinus floor, 13 mm from the sinus floor, and 15 mm from the alveolar crest were calculated on the coronal images (Fig. 2b–d).

Distance to the inferior border of the vessel

Distances to the inferior border of the vessel from the sinus floor and alveolar crest were measured on the coronal images (Fig. 2e and f). Also the distances to the inferior border of the vessel from the alveolar crest were classified into two categories: less than 15 mm from the crest and greater than 15 mm from the crest.

Vessel diameter

The vessel diameter was calculated on the coronal images (Fig. 2g). Then, the vessel diameter was divided into four categories: 1, no identification of a bony canal; 2, diameter less than 1 mm; 3, diameter of 1–2 mm; and 4, diameter greater than 2 mm (Mardinger et al. 2007).

Position of the vessel

The position of the vessel was examined through total coronal images from the rear to the front of the maxillary sinus. The vessel position was classified into three groups: Group 1 (superficial), Group 2 (intraosseous), or Group 3 (intrasinusal) (Fig. 2g).

Angle A

Angle A represents the angle between the buccal alveolar wall and the palatal alveolar wall. Angle A was divided into three groups: Group 1 (<30°), Group 2 (31–60°), or Group 3 (61°>) (Fig. 2h).

Septa

The presence or absence of septa was examined through total axial images (Fig. 2a).

Statistical analysis

The mean value and standard deviations were calculated. A significant correlation among the mean values of the variable measurement values of 7 points were tested by means of the one-way analysis of variance (ANOVA). If the differences among the 7 points were statistically significant, Duncan's method for post hoc analysis was performed. A correlation between the vessel diameter and lateral wall width was tested for statistical significance by means of the Pearson's correlation test. Differences in the variable measurement values (three values of the lateral wall width, two values of the distance to the inferior vessel border, vessel diameter, distribution of the vessel position, and angle A) according to age were tested by means of Pearson's correlation test. Differences in the variable measurement values by gender were tested using a t-test. Correlation of septa presence by age was tested using a t-test, while Pearson's chi-square test was conducted for gender. Statistical significance was set at < 0.05.

Results

“z − 15” was often out of bounds of the maxillary sinus. A very small population had a “z − 15”; therefore, the measurement values of six total coronal images were used in this study.

Table 1 shows the means of the variable measurement values (three values of lateral wall thickness, two values of the distance to the inferior vessel border, and vessel diameter) of six coronal images. There were statistically significant differences among the six coronal images (< 0.05). In lateral wall thickness, the areas of “z” in the rear had a significantly thinner mean thickness than in the front. The thickness of the lateral sinus wall on “z + 15” was the thinnest compared with other points. Three millimeters above the sinus floor, the thickness of the sinus lateral wall tended to be thick on “z − 10.” The thickness of the sinus lateral wall tended to be thick on “z − 5” at 13 mm from the sinus floor and 15 mm from the alveolar crest. In the distance to the vessel, the course of the vessel formed a concave arch, with the most inferior site in the “z” (Table 1, Fig. 3). Among the mean values of the vessel diameter from the six coronal images, “z” and “z − 5” were greater than the other points.

Figure 3.

Course of the vessel related to the mean distance from the alveolar crest (mm). “z,” as reference point, is the zygomatic process of the maxilla. “z − 5” is 5 mm of “z” in the front. “z + 5” is 5 mm of “z” in the rear. The course of the vessel formed a concave arch, with the most inferior site in the “z.” The shape of the sinus floor resembled the course of the vessel, especially the anterior area.

Table 1. Values of the anatomical structures in six coronal images (mm, mean ± SD)a
Variablesz − 10z − 5zz + 5z + 10z + 15P (ANOVA)
  1. SD, standard deviation.

  2. a

    Within the same measurement category, values with the same capital letter are not statistically different by Duncan's method for post hoc analysis.

Lateral wall thickness 3 mm from sinus floor1.94 ± 0.75a (= 141)1.80 ± 0.70ab (= 150)1.69 ± 0.84bc (= 149)1.64 ± 0.80bc (= 150)1.57 ± 0.78c (= 150)1.27 ± 0.68d (= 123)0
Lateral wall thickness 13 mm from sinus floor1.63 ± 0.68b (= 108)1.92 ± 0.67a (= 116)1.82 ± 0.78a (= 68)1.57 ± 0.68b (= 137)1.17 ± 0.55c (= 131)1.08 ± 0.43c (= 95)0
Lateral wall thickness 15 mm from alveolar crest1.79 ± 0.78a (= 103)1.79 ± 0.68a (= 142)1.70 ± 0.75a (= 137)1.33 ± 0.70b (= 142)1.24 ± 0.74b (= 135)1.19 ± 0.72b (= 106)0
Distance to inferior border of the vessel from sinus floor7.55 ± 3.94c (= 93)7.75 ± 3.55bc (= 106)7.37 ± 3.10c (= 84)8.34 ± 3.77bc (= 111)8.84 ± 4.34b (= 117)10.41 ± 4.41a (= 54)0
Distance to inferior border of the vessel from alveolar crest19.31 ± 4.93a (= 93)15.93 ± 4.50c (= 106)14.40 ± 3.99d (= 84)16 ± 4.05c (= 111)17.87 ± 3.93b (= 117)19.76 ± 3.33a (= 54)0
Vessel diameter1.16 ± 0.69b (= 93)1.31 ± 0.56ab (= 106)1.36 ± 0.57a (= 84)1.15 ± 0.48b (= 111)1.12 ± 0.53b (= 117)1.18 ± 0.49b (= 54)0.015

There were statistically significant differences between the vessel diameter and lateral wall thickness (= 0.005, 0.001, and 0.001), as shown in Table 2. As a result, it was inferred that the thicker the sinus lateral wall, the greater the vessel diameter.

Table 2. Pearson's correlation test of vessel diameter and lateral wall thickness
Vessel diameter3 mm from sinus floor13 mm from sinus floor15 mm from alveolar crest
Pearson's correlation coefficient0.2390.2920.284
P-value0.0050.0010.001

The means of the variable measurement values are shown in Table 3. The mean thickness of the lateral wall at 3 mm above the sinus floor, 13 mm above the sinus floor, and 15 mm above the alveolar crest was 1.67, 1.53, and 1.57 mm, respectively. The mean distance to the vessel inferior border from the sinus floor and from the alveolar crest was 8.25 and 17.03 mm, respectively. The mean vessel diameter was 1.18 mm. A significant correlation (= 0.042) was found between age and lateral wall thickness 13 mm from the sinus floor. Although the P-value of the vessel diameter was >0.05, it nearly showed a significant correlation with age (= 0.055).

Table 3. Pearson's correlation test for significant correlations between age and means of variable measurement values of distance (mm)
Variables N MeanSDPearson's correlation coefficientP-value
  1. SD, standard deviation.

  2. a

    Statistically significant differences in mean variable measurement values by age were tested for using Pearson's correlation test (< 0.05).

Lateral wall thickness 3 mm from sinus floor1501.670.500.0700.396
Lateral wall thickness 13 mm from sinus floor1441.530.450.1700.042a
Lateral wall thickness 15 mm from alveolar crest1481.570.580.1130.173
Distance to inferior border of the vessel from sinus floor1358.253.25−0.0610.483
Distance to inferior border of the vessel from alveolar crest13517.033.530.0620.477
Vessel diameter1351.180.450.1660.055

Table 4 presents the means of the variable measurement values by gender. There was a significant difference between men and women in the lateral wall thickness 3 mm from the sinus floor (= 0.002). Men (1.78 mm) tended to have a thicker lateral wall than women (1.52 mm) in the area 3 mm from the sinus floor. The vessel diameter was significantly different between men and women (= 0.027), with men (1.24 mm) having a larger vessel diameter than women (1.07 mm).

Table 4. Analysis for differences in means of variable measurement values by gender (mm)
VariablesGender N MeanSDP-value
  1. SD, standard deviation.

Lateral wall thickness 3 mm from sinus floorMale901.780.500.002
Female601.520.47
Lateral wall thickness 13 mm from sinus floorMale851.590.490.066
Female591.450.38
Lateral wall thickness 15 mm from alveolar crestMale901.630.630.147
Female581.480.48
Distance to inferior border of the vessel from sinus floorMale838.543.240.189
Female527.783.25
Distance to inferior border of the vessel from alveolar crestMale8317.353.470.187
Female5216.523.61
Vessel diameterMale831.240.440.027
Female521.070.43

The superficial group among the vessel position was 6.6%, the intraosseous group, 64.3%, and the intrasinusal group, 29.1%. No obvious correlation between vessel position and age (= 0.140) or gender (= 0.880) was observed. In angle A, the group of less than 30° was 4.8%, 31–60°, 42.8%, and greater than 61°, 52.4%. There was no statistically significant difference between the value of angle A and age (= 0.762) or gender (= 0.195). The prevalence of septa was 44%; no correlation was found with age (= 0.345) or gender (= 0.638).

The distance to the inferior border of the vessel from the alveolar crest less than 15 mm was observed 31% of the time, and greater than 15 mm was observed 69% of the time. Vessel diameters less than 1 mm and greater than 1 mm were observed 62.2% and 37.8% of the time, respectively.

Discussion

Alveolar bone resorption caused by tooth loss is characterized by its loss in height, which is accompanied not only by central bone repair but also by marginal or “cortical” bone resorption. It is therefore important to know whether or not resorption is due to extraction or senescence. There are two factors contributing to the loss of bone; one is related to basal bone loss due to the reinforcement of the osteoclastic activity of the sinus membrane, and the other is related to alveolar bone loss due to the disappearance of the marginal bone (Chanavaz 1990; Garg 1999). Even a slight increase in positive intra-antral pressure can cause enlargement in the volume of the maxillary sinus (Smiler et al. 1992).

Sinus elevation procedures depend greatly on fragile structures and anatomical variations. The classical sinus lift operation consists of the preparation of a top hinge door in the lateral maxillary sinus wall. This door is luxated inward and upward together with the Schneiderian membrane to a horizontal position, thus forming the new sinus bottom. The space underneath this lifted door and sinus mucosa is filled with graft material. Even though the principle of sinus elevation is simple, there are still a number of anatomical aspects and considerations connected with this type of surgery (van den Bergh et al. 2000).

In the surgical procedure, the inferior horizontal bone cut is made approximately 3 mm above the crest of the alveolar ridge. The vertical osteotomies are 10–15 mm long. When possible, vertical bone cuts are made that are long enough for the inferior aspect of the osseous window to make contact with the medial wall of the antrum when this window is reflected inward. However, anatomic restrictions often preclude this requisite. For patients with sufficient height in the lateral wall of the maxilla, the cavity created may be large enough to accommodate a graft of sufficient height. However, compromised anatomy of the lateral wall of the maxilla and the malar bone is common (Smiler et al. 1992; Smiler 1997). If the goal is to place implants 13–15 mm in length, the superior osteotomy cut should be made approximately 15 mm from the alveolar crest, and the inferior cut should be made approximately 2–3 mm from the floor of the sinus (Elian et al. 2005). In this study, the measuring area of the lateral sinus wall thickness utilized the aforementioned values referred to in these studies.

Among the means of the measurement values of the six coronal images, there were significant differences (< 0.05). In the result of Duncan's method for post hoc analysis, the thickness of the anterior area of the sinus lateral wall was significantly thicker than that of the posterior wall. Therefore, Schneiderian membrane perforation might occur while performing the antrostomy with a rotary cutting instrument, especially in the posterior area.

The anastomosis formed a concave arch, with the first molar area being the lowest point of the bony canal arch course (Traxler et al. 1999; Mardinger et al. 2007). In the present study, the course of the bony canal formed a concave arch at the most inferior site at the “z,” that was similar result to the previous studies. According to Fig. 3, the shape of the sinus floor resembled the course of the vessel, especially the anterior area.

There were statistically significant differences between the vessel diameter and lateral wall thickness (= 0.005, 0.001, and 0.001); a thicker sinus lateral wall corresponded to a larger vessel diameter. During sinus elevation with a thicker lateral wall, there is an increased risk of bleeding. According to Elian et al. (2005), an intraosseous vascular canal at the lateral wall has been identified in over 50% of examined CT images. Therefore, when the sinus lateral wall thickness is large, the risk of bleeding should be considered even if there is no intraosseous vessel in the CT images.

A vestibular extraosseous anastomosis is located at a mean distance of 23 mm from the alveolar ridge and the endosseous anastomosis at 18.9 mm from the alveolar ridge. The common vessel runs in the canal that can be easily damaged while preparing the bony window for sinus elevation. Transecting this vessel can cause minor to intense bleeding that may obscure vision and lead to perforation of the Schneiderian membrane (Solar et al. 1999). In this study, the mean distance to the inferior vessel border from the alveolar crest was 17.03 ± 3.53 mm, and similar results were published by several authors (Solar et al. 1999; Elian et al. 2005; Mardinger et al. 2007).

According to Hur et al. (2009), the shortest distance from the sinus floor to the intraosseous branch of the PSAA was 2.5 mm at the first molar region. They concluded that the intraosseous branch of the PSAA was subject to injury during the modified Caldwell-Luc approach. That study included the dentate only, and it has been shown that after tooth extraction, the sinus anatomical structure undergoes several changes as the alveolar bone shows resorption and pneumatization of the sinus (Chanavaz 1990; Ulm et al. 1995a; Garg 1999). As the implant procedures are performed on the edentulous area, it is an illogical conclusion that the vessel is subject to injury during the modified Caldwell-Luc approach.

In this study, a significant correlation (= 0.042) between age and lateral wall thickness of 13 mm from the sinus floor was observed. In other words, it is possible that the older the patient, the thicker the lateral wall on the area of 13 mm from the sinus floor.

In older, edentulous populations, the number of vessels and vessel diameter decreases, while tortuosity of the vessels increases (Ulm et al. 1995a). Other authors stated that the older the patient, the wider the diameter (Mardinger et al. 2007). Based on the results from the present study, although the P-value was >0.05, it is likely that the older the patient, the wider the diameter of the vessel (= 0.055).

T-tests revealed that the value of the lateral wall thickness 3 mm from the sinus floor between men and women differed, and the difference was statistically significant (= 0.002). The vessel diameter was also significantly different between men and women (= 0.027). Men had a greater mean lateral wall thickness 3 mm from the sinus floor and a greater vessel diameter than women. This would be an ideal explanation for men exhibiting a greater tendency for bleeding than women during the surgical procedure.

The mean vessel diameter was 1.18 ± 0.45 mm. Vessels with no identification of the bony canal or with a diameter less than 1 mm were observed 62.2% of the time, whereas vessels with a diameter greater than 1 mm were observed in 37.8% of cases. This result suggests that there is great potential for control of intraoperative bleeding. Mardinger et al. (2007) reported that the canals were over 1 mm wide in 29% of cases and therefore at risk for bleeding. These results are consistent with those of the present study. Additionally, Ella et al. (2008) reported that the overall mean diameter of the vessels was 1.20 mm, and 57.1% of patients had an artery 1–2.5 mm in diameter, indicating the probability of high-risk hemorrhage.

Herein, the vessel positions were classified into three groups: intraosseous (64.3%), intrasinusal (29.1%), and superficial (6.7%). The results correlate well with previous studies (Elian et al. 2005; Mardinger et al. 2007) that reported that an intraosseous vascular canal at the lateral wall had been identified in over 50% of the examined CT images. These results taken together suggest that it is possible to reduce the risk of bleeding by preoperative examination.

Angle A is classified into three groups. Group less than 30° is 4.8%, 31–60° is 42.8%, and more than 61° is 52.4%. There is low risk of Schneiderian membrane perforation by narrow angle.

Perforation is most likely to happen at sharp edges and ridges like the Underwood septa and at the spines (Chanavaz 1990). In the present study, the prevalence of septa related to Schneiderian membrane perforation was 44%. The presence of septa can be evaluated using the axial plane of the CT images. The reported incidence of septa varied from 16% to 33% (Underwood 1910; Ulm et al. 1995b; Krennmair et al. 1997, 1999; Velásquez-Plata et al. 2002; Kim et al. 2006), and septa were much more frequent on the left side than the right (approximate ratio of 3 to 1) (Underwood 1910). Based on this finding, it was assumed that the prevalence of septa in the current study would be higher than in the other studies because more left than right sinuses were analyzed (85 of 150 total).

It observed in this study that 69% of arteries were located more than 15 mm from the alveolar crest, whereas 31% were located less than 15 mm from the alveolar crest. This indicates that 31% of the cases in this study present the potential for intraoperative bleeding. Previous studies have produced similar results. Elian et al. (2005) reported that 80% of arteries were located more than 15 mm above the alveolar crest, therefore only 20% of the cases presented potential surgical complications.

An understanding of anatomy is essential for successful performance of fragile types of surgery. As such, preoperative radiological investigations can be very important (van den Bergh et al. 2000; Timmenga et al. 2003). Several authors have suggested that CT images can be useful for clinicians in diagnosis and treatment planning by enhancing the accuracy of the diagnostic decisions and facilitating the formulation of adequate treatment plans (Schwarz et al. 1987; Alder et al. 1995; Krennmair et al. 1997).

From the findings of this study, within the limits of high standard deviation, there are several considerations related to intraoperative complication of sinus elevation surgery. First, while performing an antrostomy with a rotary cutting instrument, it is possible to damage the vessel by intersection of cutting line and vessel in the anterior area, and the Schneiderian membrane by thin lateral wall in the posterior area. Second, when the sinus lateral wall thickness is large, the risk of bleeding should be considered even if there is no intraosseous vessel in the CT images. Third, the mean distance to the inferior vessel border from the alveolar crest was 17.03 ± 3.53 mm, and 69% of vessels were located more than 15 mm from the alveolar crest. Though the alveolar crest has been severely resorbed depending on period of extraction, the results of present study is informative guide of sinus elevation surgery because it evaluated the CBCTs images of patients for implant-supported restorations. Forth, women had a thinner mean lateral wall thickness 3 mm from the sinus floor, so the risk of membrane perforation should be considered. On the other hands, men had a greater vessel diameter than women. There is the likelihood that men have a greater tendency for bleeding than women. Fifth, it is possible that the older the patients, the wider the diameter of the vessel. One would expect this vessel to be more of a risk for a bleeding in older patients.

In patients with severe atrophy of the posterior maxilla, sinus elevations have proven successful. Nevertheless, several complications related to sinus elevation still exist. Based on the results of the present study on the utilization of cone beam CT scans for sinus elevation, the alteration of the lateral approach sinus elevation is highly recommended if complications such as membrane perforation or bleeding are expected.

Acknowledgements

Funding: The authors declare no conflicts of interests. This study was funded by the Department of Periodontology, School of Dentistry, Kyung Hee University, Seoul, Korea.

Ancillary